Battery State-of-Health Prediction-Based Clustering for Lifetime Optimization in IoT Networks

The Internet of Things (IoT) represents a pervasive system that continuously demonstrates an expanded application in various domains. The energy-efficiency problem has always been a crucial issue linked to this type of network where the system lifetime strongly depends on devices' batteries. Numerous energy-efficient networking protocols have been proposed in the literature to increase the system lifetime. However, most of the proposed approaches deal with the short-term vision of energy consumption and omit to consider the rechargeable battery degradation when evaluating the network lifetime. Indeed, the major parts of the network devices use rechargeable batteries that age and degrade over time due to several factors (temperature, voltage, charging/discharging cycle, etc.). Therefore, it is essential to promptly detect these internal and environmental degradation factors to avoid network failures. Clustering represents one of the main wireless network protocols and plays an essential role in network self organizing. In this work, we propose a novel long-term energy optimization clustering approach based on battery State of Health (SoH) prediction, called LECA_SOH. The objective is to predict the impact of cluster heads election on the rechargeable batteries SoH before applying the clustering. LECA_SOH fosters the selection of the nodes, which will less suffer from battery degradation during the future rounds, leading to extend the system lifetime. The obtained results demonstrate that the proposed clustering approach improves the network lifetime in the long term and extends the number of recharging cycles compared to the conventional energy-efficient approaches.